2354
Vol. 5
Indian Journal of Science and Technology
No. 3 (Mar 2012)
ISSN: 0974- 6846
Monitoring of induction motor temperature under unbalanced supplying by stator resistance estimation
*
Masoud Sabaghi and Hassan Feshki Farahani
Department of Electrical Engineering, Ashtian Branch, Islamic Azad University, Ashtian, Iran
sab20022003@yahoo.com
*
Abstract
Motor parameters estimations are needed for many applications. One of these parameters is stator winding resistance.
Induction motor temperature can be monitored by this parameter estimation under balanced and unbalanced condition.
For this purpose, a DC signal is injected to AC voltage. In this study, stator temperature is obtained from two methods.
One of them is measuring by construction of a temperature measurement system using sensor-installing method and
the other method is measuring temperature by Stator Winding Estimating (SWE). Each method is implemented on a
1.1kW / 50Hz / 1400rpm three-phase squirrel cage induction motor. After that, different experimental tests have been
done on the motor under balanced and unbalanced conditions and they have been evaluated. The capability of
measuring temperature from SWE method is investigated and responding of this method to motor thermal
characteristics variations and reliable thermal protection are evaluated.
Keywords: Sensor Installing, DC Injection, Temperature Monitoring, Induction Motor and Stator Winding.
Introduction
method in (Bai et al., 2010). In this paper, heat resource
Induction motors have many advantages which of motor, equivalent coefficient of heat conductivity for the
among them low cost, high reliability, low inertia and high stator coil, and coefficient of heat convection in the airtransient torque capacity can be mentioned. In these gap are analyzed. Furthermore, the temperature
motors, about 35%–45% of motor failures are caused by distribution of motor is simulated and results are
stator insulation breakdown. One of the operation, and analyzed. The influences of broken bars located at
situations where the cooling ability of the motor is different relative positions on an induction motor are
accidentally reduced (Farag et al., 1994; Hurst & presented by Ying (2010). In this investigation, a finiteHabetler, 1997).
element model of the squirrel-cage induction motor is
Many papers are presented about thermal behavioral developed. Valenzuela and Ramírez (2011) valuated
and modeling of induction motor (Mendes et al., 2008; three thermal models that are capable of detecting effects
Anwari & Hiendro 2010; Bai, 2010; Yu et al., 2010; of pulp on dissipation of operating losses to the ambient.
Valenzuela & Ramírez 2011; Wrobel et al., 2011). Thermal analysis of a segmented stator winding design is
Influence of unbalance factors on calculating total copper presented by Wrobel et al. (2011). This approach allows
losses, efficiency, power factor, input power, output for a rapid and inexpensive assessment of thermal
torque, peak currents, and de-rating factor of motor performance of the complete machine and early
operating under unbalanced voltage system is evaluated identification of design modifications needed.
in (Anwari & Hiendro, 2010).
In this paper, DC injection method is used for
Reference (Mendes et al., 2008) describes a series of estimation of stator winding resistance for temperature
experimental tests that were conducted on a 4 kW total monitoring of induction motor under unbalanced
enclosed fan cooled (TEFC) three-phase induction motor. supplying condition. The stator temperature is obtained
The tests were carried out when motor is supplied by both using two methods, which are temperature measurement
voltage source inverter and sinusoidal voltage source, with sensor installing method and DC injection method.
considering the influence of the ventilation. A Each method is implemented on a 1.1kW/ 50Hz/ 1400rpm
comprehensive transient temperature of both stator and three-phase squirrel cage induction motor. To verify DC
rotor is presented. The 3-D temperature rise of induction injection method efficiency, different tests are carried out
motor is analyzed and computed using finite element under unbalance condition.
Table 1. Resistance measurement and estimation or motor temperature using resistance
Time [Min]
Voltage [V]
Current[A]
Winding
Resistance
[Ω]
Measured
O
Temp [ c]
Estimated
O
Temp [ c]
1
12.15
4.02
5
12.34
3.98
10
12.52
3.98
15
12.57
3.99
20
12.64
3.99
25
12.61
3.97
30
12.56
3.94
35
12.65
3.94
40
12.77
3.95
45
12.86
3.95
50
12.88
3.93
55
12.90
3.93
60
12.98
3.92
3.02
3.1
3.14
3.15
3.17
3.18
3.19
3.21
3.23
3.25
3.28
3.29
3.31
30
34
37
38
40
41
42
44
45
45
50
55
60
30.2
36.6
40.4
40.8
42.2
42.9
43.9
45.7
47.6
12.86
12.88
12.90
12.98
Research article
Indian Society for Education and Environment (iSee)
“Monitoring the induction motor temperature”
http://www.indjst.org
M.Sabaghi & H.F.Farahani
Indian J.Sci.Technol.
2355
Vol. 5
Indian Journal of Science and Technology
Estimation of inductive motor temperature by measuring
its resistance
By connecting one of phases of 3-Φ squirrel-cage
induction motor windings to DC power supply and
increasing current value to 4A, winding voltage and
current can be measured. Also, a digital temperature
sensor with 1oC accuracy has been installed in this
winding to measure its temperature. A delicate
voltmeter and ampere meter are also used to
measure winding voltage and current.
Examination goal is to determine motor winding
resistance variations and to estimate its temperature
as well. Circuit voltage and current values which are
measured for one hour (each 5 min) and winding
temperature values (determined by sensor) are
listed in Table.1.
In another row of this table, estimated
temperatures are shown which are obtained by
replacing resistance in equation R T   cu R 0 which
 cu  0.004 K 1 .
Stator resistance increases due to temperature
increase (Fig. 1). Both measured and estimated
temperatures are compared (Fig. 2). The estimated
temperature values are approximately close to measured
ones. In this investigation due to using DC voltage and
current, resistance calculation becomes simple; however,
motor works by AC voltage supply. A question that may
arise here is how motor winding resistance can be
determined while the motor works by AC power
supply.The answer is that a DC component should be
added to motor voltage. As a result, similar to previous
examination, a DC current will be generated in motor.
Then by computing ratio of DC voltage to DC current,
motor resistance will be determined. For this purpose, a
DC injective system should be placed in motor basic
circuit supply. How to locate this DC injector is shown in
Fig. 3.
DC injector circuit
DC injector circuit is composed of a resistance (Rext)
and a MOSFET switch as shown in Fig. 4. If the switch is
turn on, in positive cycle, most of I2 flows through switch
and in negative cycle, all I3 flows through MOSFET
internal diode. When switch is turn off, I1 in positive and
No. 3 (Mar 2012)
ISSN: 0974- 6846
negative cycles flows through MOSFET resistance and
diode respectively, consequently, when switch is turn on,
resistance is one of motor phases is parallel with Rds,on,
Rext resistances. Because of small value of Rds,on
resistance while MOSFET is turn on , three-phase current
flows in motor.
Now, if switch is turn off, Rext in positive cycle and
diode resistance (very low) is series with one of the motor
phases. Therefore, current amplitude in positive cycle is
smaller than that in negative cycle. From current
waveform in Fig. 5, it can be realized that a DC
component have been added to it and switch voltage will
also contain DC component. Motor equivalent resistance
value can be obtained by current and voltage values. This
resistance is:
2 Vdc
Rs 
(1)
3 I dc
Total Structure of stator resistance estimator system is
practically implemented which is illustrated in Fig. 6.
System operation is divided into two parts: DC injection
and normal parts. In former case (DC injection), system is
operated in a way that processor sends turning off signal
to MOSFET driver. As it was stated before, by turning off
MOSFET, current amplitude in positive cycle will be less
than its negative amplitude due to Rext resistance. In other
words, there will be a DC component in switch voltage. At
this time, processor reads switch voltage and current by
A/D converter and after some calculations; it sends this
information to computer by RS232 port in order to
temperature monitoring process.
In latter part (normal case), turning-on signal is sent to
MOSFET by processor. When MOSFET is activated, its
voltage drop is insignificant due to subtle resistance
value. Therefore, DC injection mode does not exist in this
case. To decrease losses in switch circuit and induction
motor winding, injection time is much more less than
normal mode time. Here, normal time according to load
type is considered each 5 minutes. Also, injection time is
considered about 15 cycles of grid voltage (about 300
m/sec). For heavy loads, motor resistance sampling
should be carried out in shorter period of time (for
instance every second) because of rapidly increasing
motor temperature. In light loads, normal case is about
every 10 minutes. Different parts of circuit are described
in following sections.
Table 2. Motor resistance estimation under no load condition
Time [Min]
Estimated
Resistance
[Ω]
Sensor
O
Temp [ C]
Estimated
O
Temp [ C]
O
Error [ C]
1
5
10
15
20
25
30
35
40
45
50
55
60
2.99
3.00
3.02
3.04
3.03
3.03
3.04
3.03
3.05
3.04
3.05
3.07
3.09
25
26
26
27
28
28
29
29
30
30
30
31
31
24.5
25.6
27.3
28.2
27.7
28.1
28.6
27.9
29.2
28.7
29.3
31.3
32.5
-0.5
-0.4
+1.3
+1.2
-0.3
+0.1
+0.4
-1.1
-0.8
-1.3
-0.7
+1.3
+1.5
Research article
Indian Society for Education and Environment (iSee)
“Monitoring the induction motor temperature”
http://www.indjst.org
M.Sabaghi & H.F.Farahani
Indian J.Sci.Technol.
2356
Vol. 5
Indian Journal of Science and Technology
No. 3 (Mar 2012)
ISSN: 0974- 6846
3.4
Rext
3.3
)
m 3.2
h
o(
s
R
3.1
3
0
10
20
30
time(minute)
40
50
60
Fig.1. Stator resistance increasing due to temperature
increasing for one hours
55
Estimated
Fig.6. Structure of stator resistance estimator
Measured
50
45
e
r
u
ta
r
e
p
m 40
e
T
35
30
0
10
20
30
time(minute)
40
50
60
Fig. 2. Measured and estimated temperature of stator
Fig.7. General monitoring system of motor temperature
using sensor installation method
3.12
3.1
3.08
Fig.3. Location of DC injection circuit between motor
and supply (Lee and Habetler 2003)
) 3.06
m
h
o
( 3.04
s
R
3.02
I3
3
Rext
Is
2.98
0
I1
20
30
time(minute)
40
50
60
Fig.8. Estimation of stator resistance using DC injection
method in no-load condition
MOSFET
I2
10
33
Fig.4. DC injector circuit (Lee and Habetler 2003)
)e
d
a
rg
ti
n
e
C
(e
r
tu
a
re
p
m
e
T
i DC
P (W)
VDC
dc
‫ﻣﺪ ﺗﺰرﯾﻖ‬
‫ﻣﺪ ﻃﺒﯿﻌﻲ‬
Measured
32
Estimated
31
30
29
28
27
26
25
P (W)
24
0
Fig.5. Waveforms of VDC and iDC under DC injection mode
and normal mode
‫ﻣﺪ ﻃﺒﯿﻌﻲ‬
Research article
Indian Society for Education and Environment (iSee)
10
20
30
time(minute)
40
50
60
Fig.9. Stator temperature obtained from measured and
estimated method in no-load condition
“Monitoring the induction motor temperature”
http://www.indjst.org
M.Sabaghi & H.F.Farahani
Indian J.Sci.Technol.
2357
Vol. 5
Indian Journal of Science and Technology
3.8
]
C
[
re
u
t
a
r
e
p
m
e
T
r
o
t
ta
s
f
o
g
n
i
s
i
R
3.6
3.5
)
m 3.4
h
(o
s
R 3.3
3.2
3.1
3
0
20
40
60
time(minute)
80
100
Measured
80
70
65
60
 no-Load Balance
55
 Locked Rotor 1-OV
50
45
40
0
5
10
15
20
25
Time [Minute]
30
35
Estimated

90
)
e
d 70
ra
g
it
n
e
60
C
(
re
u
t
a
r 50
e
p
m
e
T
40

Measured
Estimated
80
30
0
20
40
60
time(minute)
80
100
120
]
C
[
e
r
tu
ra
e
p
m
e
T
r
o
t
a
t
s
f
o
g
in
is
R
Fig 11. Stator temperature obtained from measured and
estimated method under no-load and locked rotor
60
 no-Load 2-UV
 Locker Rotor 1-OV
50
40
30
0
5
10
15
20
Time [Minute]
25
30
Fig 14. Stator temperature obtained from measured and
estimated method with unbalance supplying in the increasing
and decreasing
no-Load 2interval
-UV
Measured
Estimated
60
70
20
Rising Balance & Falling Balance condition
65
40
Fig 13. Stator temperature obtained from measured and
estimated method with unbalance supplying in the
increasing interval and balance supplying decreasing
interval
Rising Unbalance[1 -OV] & Falling Unbalance[2 -UV]condition
Fig.10. Estimation of stator resistance under lockedrotor condition
90
75
35
120
ISSN: 0974- 6846
Measured
Estimated
80
3.7
]
C
[
re
u
t
a
r
e
p
m
e
T
r
o
t
a
t
s
f
o
g
n
i
s
i
R
No. 3 (Mar 2012)
Rising Unbalance[1-OV] & Falling Balance condition
85
55
55
50
 no-Load
50
]
C
[
e
r
u
t
a
r
e
p
m
e
T
r
o
ta
t
S
f
o
45
 Locker Rotor
40
35
g
in
s
i
R
30
25
0
5
10
15
20
Time [Minute]
25
45
40
35
30
1.93%
1.93%
3.83%
3.83%
25
30
20
Fig 12. Stator temperature obtained from measured and
estimated method with balance supplying in the increasing interval
and balance supplying decreasing interval
Locked Rotor
0
5
10
15
20
25
Time [Minute]
30
[measured]
[Estimated]
[measured]
[Estimated]
35
40
Fig 15. Stator temperature obtained from measured and
estimated method in no-load with different unbalance
supplying [2φ-UV].
1-OV
80
]
C

[
e
r
tu
a
r
e
p
m
e
T
r
o
ta
t
fS
o
g
n
i
s
i
R
70
60
50
Balance [measured]
Balance [Estimated]
5.91% [measured]
5.91% [Estimated]
9.94% [measured]
9.94% [Estimated]
40
30
20
0
1
2
3
4
5
6
Time [Minute]
7
8
9
10
11
Fig 16. Stator temperature obtained from measured and
estimated method in no-load with different unbalance
supplying [1φ-OV]
Research article
Indian Society for Education and Environment (iSee)
“Monitoring the induction motor temperature”
http://www.indjst.org
M.Sabaghi & H.F.Farahani
Indian J.Sci.Technol.
2358
Vol. 5
Indian Journal of Science and Technology
Experimental results
After implementation this system, some experiments
have been carried out on squirrel cage induction motor to
verify system efficient operation. Digital sensors of
STM160 are installed in winding of each phase. Then,
temperature estimated by this system is compared with
that estimated by installed sensors in stator; Resistance
of each phase is 3.02 Ω at room temperature. This
system is shown in Fig. 7 which consists of induction
motor, DC generator and resistance estimator system.
No-load experiment
No. 3 (Mar 2012)
ISSN: 0974- 6846
A, 380V 1.1 kW and 140 rpm. These tested
characteristics are listed in Table. 3. Unbalancing
percentage has been defined in various ways among
which definitions of NEMA and IEC are most important. In
NEMA standard, unbalancing percentage is defined as
follow:
Voltage Unbalance = Vmax  100 (2)
Vavg
Where Vavg is average voltage and ΔVmax is maximum
difference between line voltage and average voltage.
Voltage unbalanced definition presented by IEC standard
is as follow:
In this experiment, phase voltage is 229 V/ 50 Hz. NoV
load current for each phase is considered 0.86 and
VUF  n  100 (3)
angular speed is determined about 2900 rpm. Due to low
Vp
thermal losses and consequently slow temperature
Where Vn is negative sequence voltage and Vp is
increasing process, sampling has been done every 5 min.
positive sequence voltage. Different tests have been
Estimated resistance by switch and calculated carried out to investigate voltage unbalancing effects on
temperature by sensor are listed in Table. 2. Resistance motor temperature.
increasing curve is plotted in Fig. 8. Estimated and
calculated temperature values have been compared and First test
results are illustrated in Fig. 9. As it is obvious from
Stator temperature using both methods under balance
obtained results, the maximum error of measured values supply condition is shown in Fig. 12. Regarding of this
o
is 1.5 C.
figure, in temperature rising interval motor works under
Locked-rotor experiment
locked-rotor condition and stator temperature starts to
This experiment has been considered in locked-rotor increase. After 12 minutes, stator temperature reaches to
condition. Motor phase voltage is 40 V and its phase 60 0C.
current is 3 A. Resistance and sensor temperatures
After this time, motor is fed from balance supply in nosampling is done every 5 minutes. After 80 minutes, to load case and stator temperature begins to decrease.
cool the motor, it is changed to no-load condition.
Also, this figure shows that motor temperature can be
In this case, motor current is calculated 0.87 A. In Fig. determined relatively accurate by estimation of stator
10, estimated resistance curve and in Fig. 11, comparing winding resistance.
results of measured and estimated temperatures are
shown. According to these figures, up to 80 minutes, Second test
because of high motor current value, resistance
In this test, motor characteristics is 1Φ-OV unbalance
estimation is better.
supply
with
unbalance
percentage
of
9.94%
Investigation of voltage unbalancing effects on motor (VUF=9.94%). Motor starts to work in locked-rotor
temperature
condition and its temperature is increasing and after 15
Different experiments have been accomplished to min, it reaches to 82oC. After that, motor works in balance
investigate different parameters effects on motor and no-load condition and its temperature is
increasing temperature which some of them are approximately equal to calculated temperature. For this
mentioned in this paper.
condition, temperature variation is plotted in Fig. 13.
Experiments are in two balanced and unbalanced Third test
voltage supplying conditions and include locked-rotor and
In this test, motor works under these conditions:
no-load conditions. In locked-rotor case, motor drown unbalance power supply, 1Φ-OV with VUF= 5/91% and
nominal current from power supply. Nominal current is 3
Table 3. The property of supplying voltage and motor operational condition (LR: Locked Rotor and nL: no-Load)
case
balance
balance
1φ- OV
2φ- UV
2φ- UV
1φ- UV
1φ- OV
balance
2φ- UV
2φ- OV
VCA [Volt]
88
370
90
351
351
86
100
365
326
405
VBC [Volt]
88
370
106
370
370
46
131
365
365
372
Research article
Indian Society for Education and Environment (iSee)
VAB [Volt]
88
370
88.5
348
348
86
97
365
326
411
Vp
88
370
94.83
356.33
356.33
72.67
109.33
365
339
396
Vn
0
0
5.6
6.89
6.89
13.33
10.87
0
13
12.12
V0
0
0
5.6
6.89
6.89
13.33
10.87
0
13
12.12
“Monitoring the induction motor temperature”
http://www.indjst.org
VUF( % )
0
0
5.91
1.93
1.93
18.35
9.94
0
3.83
3.06
Temperature
Rising
Falling
Rising
Falling
Rising
Rising
Rising
Falling
Rising
Rising
condition
LR
nL
LR
nL
nL
LR
LR
nL
nL
nL
M.Sabaghi & H.F.Farahani
Indian J.Sci.Technol.
2359
Vol. 5
Indian Journal of Science and Technology
locked-rotor condition. At first, motor temperature is
o
increasing and it reaches to 83 C after 12 m. after this
time, motor starts working under unbalance, 2Φ-UV with
VUF=1.93% in no-load condition. Motor temperature is
decreasing which is shown in Fig. 14.
Investigation of voltage unbalance impact on motor
temperature decreasing
To investigate voltage unbalance influence on motor
temperature, different tests have been conducted which
are explained in following parts:
Motor temperature increasing in 2Φ-UV for no load
condition
In this experiment, two unbalancing cases of VUF
=1.93% and 3.83% and 2Φ-UV have been tested in noload condition.
According to results shown in Table 3, for each
unbalancing, negative and zero sequence amplitude is
equal to 6.89 V and 13V respectively which are expected
that these components cause motor temperature
increasing. For this test, motor temperature obtained from
both methods and for each unbalancing is plotted in Fig.
15. It shows that estimated results are approximately
equal to calculated values.
Unbalance 1Φ-OV on motor temperature increasing in
locked rotor condition
In this investigation, three cases of balance, VUF =
5.91% and VUF= 9.44 % in 1Φ-OV for a motor are
considered in locked rotor condition. As it is shown in Fig.
16, by increasing VUF, negative and zero sequence
amplitude are increased which leads to motor
temperature rising
Conclusions
In this paper, the stator temperature has been
measured by estimation of stator winding resistance of
induction motor under unbalance supplying condition. For
this purpose, a DC signal injected to AC voltage and the
resistance of stator winding has been estimated under
balanced and unbalanced conditions. The stator
temperature has been obtained by construction of a
temperature measurement system with sensor installing
method. Each method has been implemented on a
1.1kW/ 50Hz/ 1400rpm three-phase squirrel cage
induction motor and experimental results have shown that
by using DC injection method, stator winding resistance
(temperature) can be estimated with high accuracy. Also,
experimental results have indicated that increase of
motor temperature depends on the positive, negative and
zero sequence voltage. In the high percent of
unbalancing, the motor temperature increases due to
positive voltage sequence reduction.
Research article
Indian Society for Education and Environment (iSee)
No. 3 (Mar 2012)
ISSN: 0974- 6846
References
1. Anwari M and A Hiendro (2010) New Unbalance
Factor for Estimating Performance of a Three-Phase
Induction Motor With Under- and Overvoltage
Unbalance. Energy Conversion, IEEE transactions.
25, (3)619-625
2. Bai B, Yu Q et al. (2010) 3-D Thermal analysis and
computation of flameproof induction motor. Power &
Energy Engin. Conf. (APPEEC). 1-4.
3. Farag SF, Bartheld RG et al. (1994) Electronically
enhanced low voltage motor protection and control.
IEEE Trans. Industrial. Appl. 30, 776-784.
4. Hurst KD and Habetler TG (1997) Thermal monitoring
and parameter tuning scheme for induction machines.
In: Conf. Rec. IEEE-IAS Annu. Meeting 1, 136-142.
5. Lee SB and Habetler TG (2003) An online stator
winding
resistance
estimation
technique
for
temperature monitoring of line- connected induction
machines. IEEE Trans. Industrial. Appl. 39 (3).
6. Mendes AMS, XM LF et al (2008) Thermal evaluation
of TEFC three-phase induction motors under different
supply frequencies. Electrical Machines, ICEM 2008.
18th Int. Conf. 1-6.
7. Valenzuela MA and Ramírez G (2011) Thermal
models for online detection of pulp obstructing the
cooling system of TEFC induction motors in pulp area.
IEEE Trans. Industrial. Appl. 47, 719-729.
8. Wrobel RP, Mellor H et al. (2011) Thermal modeling of
a segmented stator winding design. IEEE Trans.
Industrial. Appl. 47(5), 2023-2030.
9. Ying X (2010) Performance evaluation and thermal
fields analysis of induction motor with broken rotor
bars located at different relative positions. IEEE
Transact. Magnetics. 46, 1243-1250.
“Monitoring the induction motor temperature”
http://www.indjst.org
M.Sabaghi & H.F.Farahani
Indian J.Sci.Technol.